Design takes the driver’s seat

Louise Ward
British Safety Council
11 May 2017

Throughout history, human beings have sought opportunities to increase value, efficiency and productivity while reducing the requirement for effort and interaction.

From the development of the earliest hand tools, through water power, horse power, steam, combustion and then electric engines, to computers, the internet and robotics, this objective has driven progress throughout the world.

Now we are reaching another key milestone in the evolution of technology. Machinery and equipment is becoming fully autonomous, requiring little or no human interaction at the point of work.

So, does this render us obsolete? Is a life of leisure awaiting us all as the machines get on with the business of running our world? I think not, but I do think that the role we play is changing. Rather than being involved at the point of work, we will instead be involved in enabling; writing the algorithms, setting the objectives, specifying the output and controlling the end to end process. This is a fundamental change, but also offers an important opportunity to design management of health and safety into the systems of the future.

There is already talk of the complexities of legal liability for incidents involving autonomous machines. For example, who should be held accountable for a road accident involving an autonomous vehicle?

Not the occupant, they have no control over the operation. Who then? The organisation that wrote the control system? The owner of the equipment? The answers aren’t yet defined, but rather than focusing on just the reactive element, surely the objective must be to apply the knowledge and learning gained over the last 100 or more years to design out the factors that we know can lead to accidents.

Technology is evolving at an amazing pace, so there is a need to act quickly if we are to reap the benefits of this unique opportunity to design out risk and design in health and safety.

For many years we have understood that human involvement introduces an element of uncertainty to a system or process. People think and act independently, and are therefore inherently unpredictable. However, if the level of human involvement at the point of work is reducing, then there is a real opportunity to address this issue by applying intelligence and allowing systems to learn and evolve to reduce risk.

We have already seen some elements of this in the automotive industry, where systems have been introduced to automatically apply the brakes on encountering an obstruction; limit the speed of the vehicle based on the conditions; automatically switch on headlights and windscreen wipers where required; indeed, some cars now even park themselves! All of these systems have been designed to address key causes of accidents, by reducing the need for a person to make decisions and take action, instead using data from monitoring and sensing equipment to identify a hazardous situation and apply the appropriate control measure based on objective learning from historical data. The results have been very positive in terms of incident reduction, and the systems are now being fitted to an ever increasing range of vehicles.

There is a need to manage expectations though. Things will still go wrong, especially in the early stages of development. Indeed, trials of autonomous vehicles have already been halted several times following accidents that have attracted significant media attention. However the incidence rate is significantly lower than that for standard vehicles and learning is being directly applied to continuously improve the safety and performance of the system. Nobody is advocating personal sacrifice for the sake of progress though! Small scale trials allow this learning to develop while managing the risk of incidents and allowing people to develop confidence in the system.

Whatever your views are about the advance of technology, there is no doubt that the man-machine interface is changing fundamentally.

I believe that there is an opportunity to achieve a real step change in health and safety, if we act quickly to apply historical data; design out risk; and design health and safety into the systems of the future.

Collision and injury criteria when working with robots: read HSE research here